WO2005059588A1

WO2005059588A1 - Radar

Info

Publication number: WO2005059588A1
Application number: PCT/JP2004/015903
Authority: WO
Inventors: Motoi Nakanishi
Original assignee: Murata Manufacturing Co., Ltd.
Priority date: 2003-12-16
Filing date: 2004-10-27
Publication date: 2005-06-30
Also published as: DE112004002458T5; US20070103360A1; JP4120679B2; CN1894595A; CN100478702C; JPWO2005059588A1; US7339517B2

Abstract

A radar, wherein a first threshold (TH1) is determined by the intensity of background noise, thresholds (TH21), (TH22), and (TH23) are set in specified frequency zones near peak positions in the skirt portions of peaks (P1), (P2), and (P3), the higher value of these thresholds is adopted as a final threshold for each FET bin, and a peak exceeding that threshold is regarded as a target peak and detected. Thus, target detecting accuracy can be increased by securely detecting the target peak resulting from a reflected wave from a target included in the frequency spectrum of beat signals.

Description

Specification

Radar

Technical field

The present invention relates to a device for detecting a target using radio waves, and more particularly to a radar for detecting a target based on a frequency spectrum of a beat signal of a transmission signal and a reception signal. Background art

Conventionally, an FM-CW radar using millimeter wave radio waves has been developed as an on-vehicle radar. That is, FM modulation is performed in a predetermined frequency range around a predetermined frequency to take a beat between the transmission signal and the reception signal, and the beat frequency and the transmission signal frequency decrease while the transmission signal frequency of the beat signal increases. The distance to the target and the relative speed to the target are determined by identifying the beat frequency during the movement.

[0003] In order to identify the upbeat signal and the downbeat signal, the frequency spectrum of the beat signal is obtained, and a predetermined threshold value is set to separate the signal component and the noise component.

However, since the peak of the beat signal appearing in the frequency spectrum changes due to various factors, simply setting a certain threshold value does not make it possible to separate the signal and the noise. Therefore, for example, Patent Literature 1 discloses a vehicle that determines the vehicle type of a vehicle that is a detection target and sets a threshold value according to the vehicle type!

[0005] Further, when the rearmost vehicle suddenly appears in front of the own vehicle, the threshold is set based on the peak having the maximum power in response to a change in the maximum power of the frequency spectrum. Patent Document 2 is disclosed.

[0006] Patent Literature 3 discloses setting a threshold based on the power of a plurality of peaks in response to the peak becoming smaller in accordance with the distance to the target.

[0007] Patent Document 4 discloses an apparatus that removes a peak of a virtual image by increasing a threshold value at a frequency of the virtual image generated by a harmonic, a switching frequency, or the like. Patent Document 1: JP-A-6-214015

Patent document 2: JP-A-7-311260 Patent Document 3: Japanese Patent Application Laid-Open No. Hei 4 318700

Patent Document 4: JP-A-11-344560

Disclosure of the invention

Problems to be solved by the invention

[0008] In the radar disclosed in Patent Document 1, even if the vehicle type can be determined, the actual signal strength changes according to the situation, and therefore, the noise cannot be accurately removed.

[0009] In the radars disclosed in Patent Documents 2 and 3, unless a threshold value is set in a value 'frequency range corresponding to each peak, a target with a small answer that can be originally detected will be lost. .

[0010] Furthermore, the radar disclosed in Patent Document 4 has a problem that only noise that can be assumed at an appearance position can be removed in advance.

[0011] Therefore, an object of the present invention is to detect a spectrum peak caused by a reflected wave of a target force contained in a frequency spectrum based on a frequency spectrum of a beat signal, thereby more reliably detecting the spectrum peak. An object of the present invention is to provide a radar with improved target detection accuracy.

Means for solving the problem

[0012] The present invention provides a means for transmitting a frequency-modulated transmission signal and generating a beat signal including a frequency component of a difference between a frequency of a reflection signal from a target of the transmission signal and a frequency of the transmission signal. A radar for obtaining a frequency spectrum of the beat signal; and a means for obtaining a peak frequency of a peak appearing in the frequency spectrum, wherein the radar is configured to detect a target based on the peak frequency.

Based on the background noise intensity or the reflected signal intensity of a target with a predetermined reflection cross section, determine the first threshold value and the first threshold value that appears in the frequency spectrum. Means for determining a second threshold value in the vicinity of a predetermined frequency of each peak in accordance with the intensity of each peak, and extracting peaks exceeding the second threshold value. It is characterized by that.

[0013] Also, the present invention is characterized in that the second threshold value is increased in a predetermined band of a foot portion according to the spread of a peak generated by multiplying a beat signal by a window function in the frequency axis direction. Features. [0014] Further, the present invention is characterized in that the value of the second threshold is increased in a predetermined band in accordance with the spread of a peak generated by the CZN characteristic of the oscillator that generates the transmission signal in the frequency axis direction. .

[0015] Further, the present invention is characterized in that the second threshold value is set so as to gradually decrease in accordance with the intensity of the peak in a vertical direction with the frequency of the peak as a center. .

[0016] Further, the present invention is characterized in that the second threshold value is set so as to exceed the intensity of a side band that appears together with a peak due to a modulation component superimposed on a beat signal.

[0017] Further, according to the present invention, the means for extracting the peaks determines the second threshold value in the descending order of the peak value among the plurality of peaks exceeding the first threshold value, The second feature of the present invention is to extract peaks exceeding all values.

The invention's effect

When a beat signal is multiplied by a window function, a protrusion (hereinafter simply referred to as “peak”) based on a signal component appearing in the frequency spectrum is centered on the frequency of the peak (hereinafter referred to as “peak frequency”). And spreads in the frequency axis direction. According to the present invention, the first threshold is determined based on the intensity of knock ground noise or the reflected signal intensity of a target having a predetermined reflection cross-section, and the first threshold appearing in the frequency spectrum is determined. About multiple peaks

According to the intensity of each peak, the second threshold value, whose value is high at frequencies near the peak and low at frequencies near the peak, is set. Can be removed by a second threshold value, and a peak appears, and noise in the frequency domain can be removed by a first threshold value. In this way, only the original peak (hereinafter, referred to as “target peak”) that occurs in the frequency spectrum of the beat signal due to the reflected wave from the target can be correctly detected.

As the second threshold value, the window function is applied if the value is increased in a predetermined band of the tail portion in accordance with the spread of the peak in the frequency axis direction caused by multiplying the beat signal by the window function. As a result, the peak appearing due to the superimposition of random noise on the base of the original peak is not erroneously detected.

[0020] As the second threshold, a peak generated by the CZN characteristic of the oscillator that generates the transmission signal is used. If the value is increased in a predetermined band according to the spread of the peak in the frequency axis direction, a peak that appears due to random noise superimposed on the base of the original peak due to the application of the window function may be erroneously detected. Absent.

[0021] If the second threshold value is determined so as to gradually decrease in a vertical direction with the frequency of the peak as a center in accordance with the intensity of the peak, the frequency becomes closer to the original peak. When the signal intensity is high, it is possible to reliably detect only the original peak without detecting the peak caused by the random noise superimposed thereon according to the shape of the frequency spectrum.

[0022] If the second threshold value is set so as to exceed the intensity of the side band appearing together with the peak due to the modulation component superimposed on the beat signal, the side band of the original peak is erroneously detected as the peak. I can't.

According to the present invention, among the plurality of peaks exceeding the first threshold value, the second threshold value is determined in order from the one having the largest peak value, and the second threshold value of each peak is determined. By extracting peaks that exceed all points, the original peaks can be detected at high speed with a small amount of computation, and the target detection speed increases.

Brief Description of Drawings

FIG. 1 is a block diagram showing a configuration of a radar according to a first embodiment.

FIG. 2 is a diagram illustrating an example of a frequency change of a transmission signal and a reception signal that changes depending on a distance to a target of the radar and a relative speed of the target.

FIG. 3 is a flowchart showing a processing procedure for detecting a distance and a relative speed.

FIG. 4 is a diagram showing the relationship between the frequency spectrum of the window function and the spread of the tail of the peak.

FIG. 5 is a diagram showing a relationship between knock ground noise and a threshold value determined thereby, and a relationship between a reflection signal intensity of a target having a predetermined reflection cross-sectional area and a threshold value determined based thereon.

FIG. 6 is a diagram illustrating an example of a noise peak generated by superposition of noise in a foot portion near a peak.

FIG. 7 is a diagram showing an example of a threshold value set in a foot portion near a peak. FIG. 8 is a diagram showing a relationship between a threshold determined based on knock ground noise and a detected peak.

FIG. 9 is a diagram showing an example of a finally determined threshold line.

FIG. 10 is a flowchart showing a processing procedure for peak frequency detection.

FIG. 11 is a diagram illustrating an example of a noise peak generated in a foot portion near a peak due to the CZN characteristic of the oscillator in the radar according to the second embodiment.

FIG. 12 is a diagram showing a setting example of threshold values in the radar.

FIG. 13 is a diagram showing a setting example of a threshold value in the radar according to the third embodiment.

FIG. 14 is a diagram showing an example of a spectrum when an AM modulation component is present in a beat signal in the radar according to the fourth embodiment.

FIG. 15 is a diagram showing an example of a threshold value set according to sideband noise included in the beat signal.

Explanation of symbols

[0025] 17-DSP

ADC—AD converter

DAC — DA converter

VCO-voltage controlled oscillator

BEST MODE FOR CARRYING OUT THE INVENTION

The configuration of the radar according to the first embodiment will be sequentially described with reference to FIGS.

FIG. 1 is a block diagram showing the overall configuration of the radar. The transmission wave modulator 16 sequentially outputs digital data of the modulation signal to the DA converter 14. VCOl changes the oscillation frequency according to the control voltage output from the DA converter 14. As a result, the oscillation frequency of the VC Ol is FM modulated continuously in a triangular waveform. The isolator 2 transmits the oscillating signal of the VCOl to the force bra 3 and prevents the reflected signal from entering the VCOl. The coupler 3 transmits the signal that has passed through the isolator 2 to the circulator 4 and also provides a part of the transmission signal to the mixer 6 as a local signal Lo at a predetermined distribution ratio. Circuit circulator 4 transmits a transmission signal to antenna 5 side and supplies a reception signal from antenna 5 to mixer 6. Antenna 5 transmits VCOl FM modulated continuous wave transmit signal Then, a reflected signal from the same direction is received. In addition, the direction of the beam is periodically changed over the detection angle range.

[0027] The mixer 6 mixes the local signal Lo from the coupler 3 and the received signal from the circulator 4, and outputs an intermediate frequency signal IF. IF amplifier circuit 7 amplifies the intermediate frequency signal with a predetermined amplification factor according to the frequency determined by the distance. The AD converter 8 converts the voltage signal into a sampling data sequence and supplies the same to the DSP 17. The DSP 17 temporarily accumulates the digital data converted by the AD converter 8 for at least one scan (for a plurality of beam scans within a predetermined detection angle range), and executes a process described later to set a target around the antenna. The direction of the target, the distance to the target, and the relative speed of the target to the antenna are calculated.

[0028] In the above-described DSP 17, the DC removing unit 9 obtains an average value of a predetermined sampling section to be processed by the subsequent FFT in the sampling data sequence obtained by the AD converter 8. Since this average value is equal to the DC component obtained by FFT (Fast Fourier Transform), the arithmetic process of subtracting the average value from each data of all sampling intervals is performed before the FFT operation process. DC component is removed.

The window function processing unit 15 cuts out data from the data from which the DC component has been removed by the DC removing unit 9 using a window function having a predetermined shape. With this window function extraction, errors due to the truncation that occur when the time waveform is extracted into a finite sampling section and the FFT operation is performed are suppressed. For example, window function processing such as a Nodjung window 'Humming window' Blackman-Harris window is performed.

[0030] The FFT operation unit 11 analyzes the frequency component of the data of the sampling section to which the window function has been applied.

[0031] The peak detection unit 12 detects, as a peak frequency, a frequency of a signal having an intensity exceeding a predetermined threshold value in the frequency spectrum.

[0032] The distance / speed calculating unit 13 calculates the distance from the detected peak frequency to the target and the relative speed.

FIG. 2 shows an example of a shift in frequency change between the transmission signal TXS and the reception signal RXS due to the distance to the target and the relative speed. Transmission signal TXS is centered on center frequency fo Is a signal frequency-modulated in a triangular wave. The frequency difference between the transmission signal TXS and the reception signal RXS when the frequency of the transmission signal TXS rises is the upbeat frequency fBU, and the frequency difference between the transmission signal TXS and the reception signal RXS when the frequency of the transmission signal TXS falls is down. Beat frequency fBD. The time axis difference (time difference) between the triangular wave of the transmission signal TXS and the reception signal RXS corresponds to the round trip time of the radio wave to the antenna target. Also, the shift on the frequency axis between the transmission signal TXS and the reception signal RXS is the Doppler shift amount, which is caused by the relative speed of the target with respect to the antenna. The value of the upbeat fBU and the value of the downbeat fBD change depending on the time difference and the amount of the Dobler shift. Conversely, by detecting the frequency of the upbeat and the downbeat, the distance from the radar to the target and the relative speed of the target to the radar are calculated.

FIG. 3 is a flowchart showing a processing procedure of the DSP 17. First, AD converter

8 (S1), multiply by the weighting coefficient of the Hayung window (S2), perform FFT operation (S3), and take the log of the sum of squares of the real part and imaginary part at each discrete frequency to obtain the power spectrum (hereinafter , Simply referred to as “frequency spectrum”) (S4).

Subsequently, a plurality of peaks appearing in the frequency spectrum are detected, a target peak is extracted from the peaks, and the peak frequency is obtained (S5).

[0036] The above processing is performed in order for the uplink modulation section and the downlink modulation section of the transmission frequency. Also, a combination (pairing) of the peak frequencies of the plurality of protrusions extracted in the up-modulation section and the peak frequencies of the plurality of protrusions extracted in the down-modulation section is performed (S6). That is, the peak frequencies of the protruding portions caused by the same target are paired. Then, the relative distance and the relative speed of the peak frequency target are calculated (S7).

[0037] When a discrete frequency spectrum such as FFT is obtained from a beat signal, the effect of the discontinuity of the signal is suppressed by multiplying a sample of the extracted signal by a window function.

FIG. 4 is a diagram illustrating an example of signal processing by which a window function is multiplied and the resulting frequency spectrum. Here, (A) shows a data sequence from which the DC removal has been performed in a time waveform. By applying a predetermined window function shown in (B) to this data sequence, a data sequence of a fixed number of data (for example, 1024 data) is obtained as shown in (C). By performing an FFT operation on the data sequence to which this window function has been applied, a discrete frequency spectrum as shown in (D) is obtained. In FIG. 4D, circles indicate signal strength (power) at each discrete frequency. The solid line is a continuous spectrum of the window function shown in FIG. Since the frequency spectrum of the beat signal to which the window function has been applied in this way is a convolution of the beat signal and the window function, the spectrum expands in the frequency axis direction according to the spectrum of the window function, and a tail portion is formed in the spectrum.

FIG. 5 shows two examples of setting a threshold to extract the frequency spectrum force target peak of the beat signal! /. Fig. 5 (A) is based on background noise! / ヽ When setting the value, it shows the relationship between the knock ground noise and the threshold value set based on it. When the value is set based on the background noise in this way, the value is determined so that the probability that the knock ground noise exceeds the threshold value! / ヽ becomes sufficiently small. This probability is determined by the average value of the background noise and its variance. In (A) of Fig. 5, BN indicates the instantaneous value of the background noise, BNm indicates the average value of the background noise, and TH1 indicates the threshold value. The horizontal axis is time (elapsed time) and the vertical axis is signal strength.

FIG. 5B shows a case where the threshold value is set based on the reflection signal intensity of a target having a predetermined reflection cross section, and the reflection signal intensity of a target having a predetermined reflection cross section is set. And the thresholds set based on them. Here, the horizontal axis is the distance [m] to the target, and the vertical axis is the received signal strength (log memory) with the peak value being OdB. S is the theoretical value of the signal intensity of lOdBsm (received signal intensity when the reflected signal intensity is 0 when the object force of the radar reflection cross section is 10m2), and THO is based on this theoretical value. The threshold is lowered by a level that takes into account scintillation. In this way, the received signal strength decreases as the signal is reflected from a distance, and the value is changed accordingly.

However, when the received signal strength is high, around the target peak, the base of the peak is reflected by the target having the above-mentioned threshold value TH1 determined based on the knock ground noise or the predetermined reflection cross-sectional area. The above threshold value THO, which is set based on the signal strength, may exceed the threshold value THO. Peaks (hereinafter referred to as “noise peaks”) may be erroneously detected as peaks due to force signals.

[0042] Therefore, first, the spread of the peak by the window function and the upper and lower limits of the fluctuation of the noise intensity assumed from the variance of the knock ground noise are obtained. If the vertical width exceeds a predetermined amount of change in intensity change defined as a condition for peak (projection) detection, a projection due to noise may be detected as a peak. In other words, in each range bin (each frequency range bin according to the FFT frequency resolution) other than the target peak generated by multiplying by the window function, the intensity of a certain range bin is the upper limit of the upper and lower widths, and If the intensity of the adjacent range bins on both sides is the lower limit of the vertical width, the certain range bin is erroneously detected as a small but peak.

Therefore, in the frequency range where such erroneous detection may occur, the peak of the intensity does not exceed the upper limit of the sum of the swelling of the peak of the window function due to the window function and the noise intensity. Process as if it were not considered a signal.

[0044] To this end, apart from the threshold TH1 determined based on the knock ground noise, a value exceeding the upper limit of the sum of the noise and the window function spectrum is set, and a value TH2 is set. Those that exceed the target are detected as reflected signals of the target force.

[0045] The manner in which the tail of the frequency spectrum spreads differs depending on the type of window function. Also, the case where the target peak position of the frequency spectrum coincides with the position of the FFT range bin and the position is shifted (for example, when the FFT frequency resolution is 1 kHz, when the beat frequency exists at a frequency that is not an integral multiple of 1 kHz) It should be noted that there is a difference in the spread of the spectrum base. In consideration of these points, the case where the base spreads the most due to the relationship between the target peak position and the spread of the spectrum is considered as a reference.

FIG. 6 shows the spectrum tail by the window function, in particular, the appearance of noise peaks caused by the addition of noise to this spectrum! /.

FIG. 6A shows the shape of the spread of the tail near the peak by the window function. Also, (B) shows the spectrum that appears as a result of combining the noise with the spread of the tail near the peak by this window function. In these figures, the spectrum SPO with a backslash indicates that the target peak position of the frequency spectrum coincides with the position of the FFT range bin. , SP1 shows the case where it is shifted by a half range. P is the target peak and NP is the noise peak. If these noise peaks NP exceed the threshold value TH1 determined based on the background noise, these noise peaks NP are erroneously detected as target peaks.

[0048] Therefore, a threshold value different from the above threshold value (first threshold value) TH1 is determined for the tail portion of the peak.

Fig. 7 shows an example of the fluctuation width of the noise intensity near the peak due to the spread of the window function spectrum and the inclusion of noise. In (A), C is the theoretical value, U is the upper limit level due to noise contamination, and D is the lower limit level due to noise contamination. Therefore, as shown in Fig. 7 (B), the second threshold TH2, which is a value higher than the upper limit level considering noise contamination, is set.

[0049] It is still unclear which of the peaks appearing in the frequency spectrum of the beat signal is the power peak of the target peak, so that which peak! It is necessary to determine whether to set the above threshold TH2 for the part. Therefore, it is performed as follows.

FIG. 8 shows an example of a frequency spectrum including a plurality of peaks. Here, the waveform SP is a frequency spectrum, and the straight line TH1 is a threshold value set based on the average and variance of the background noise of the spectrum. This background noise is a knock ground noise included in the beat signal in a state where the reflected signal from the target is not received. The threshold value TH1 is determined in advance so that the probability that the noise is greater than the value is a sufficiently small probability.

In FIG. 8 (B), circled positions indicate peak positions where the change in signal intensity with respect to the frequency change forms a mountain shape in a range exceeding the threshold value TH1. If all the peak positions within the range exceeding the threshold value TH1 are regarded as true peaks, the peak indicated by the circle is also detected as a target peak. Therefore, a plurality of peaks exceeding the threshold value TH1 are detected, and the threshold value TH2 is set in descending order of the peak value. FIG. 9 shows an example of this. Here, the value TH1 is a threshold based on the background noise described above! /, Value TH22 is a threshold (threshold line) determined based on the peak value of peak P2, and TH21 is a threshold determined based on the peak value of peak P1. Similarly, TH23 is a threshold value determined based on the peak value of the peak P3. The threshold value line shown by the solid line in FIG. 9 is determined by using the side of the plurality of threshold values having the larger value. Therefore, a peak exceeding this threshold line is detected as a target peak.

FIG. 10 shows a processing procedure of peak frequency detection corresponding to step S5 in FIG.

First, a threshold value TH1 is obtained from the average value and variance of the knock ground noise, and a peak exceeding the value TH1 is extracted from a plurality of already detected peaks (SI1 → S12). Subsequently, a peak having the maximum peak value is detected from the peaks, and a threshold value (TH22 in the example of FIG. 9) is set based on the peak value (S13). Then, the presence or absence of a peak exceeding the threshold value (TH22) is determined. (In the example of Fig. 9, TH21)

[0054] Thereafter, the same processing is repeated, and threshold values are set for all peaks exceeding a plurality of threshold values sequentially determined (S14 → S15 → S13 → .- ·) ο Peaks exceeding multiple thresholds (PI, Ρ2, Ρ3 in the example shown in Fig. 9) are regarded as target peaks (S16).

[0055] The above-described processing is based on the threshold value ΤΗ2 set at each foot portion based on the peak value of each peak and the threshold value TH1 obtained from the average value and variance of the background noise. This is equivalent to adopting a value higher or lower than the value in the range bin as a value, and processing the peak as a noise peak. However, as described above, it is more efficient to set the threshold value in the tail part in the descending order of the peak value, so that the noise peak can be eliminated efficiently.

[0056] It should be noted that the threshold value set based on the peak value of the detected peak and the value しきい値 2 and the threshold value TH1 obtained from the average value and the variance of the knock ground noise are used as shown in FIG. 5 (Β). As described above, the threshold value した set based on the reflection intensity from the target having the predetermined reflection cross section is Of these, the range of the value may be determined in each range bin, and a process for adopting the value may be performed.

Next, a radar according to the second embodiment will be described with reference to FIGS. 11 and 12.

With respect to the FM-CW radar, when the CZN characteristic of the oscillator that generates the transmission signal and the local signal deteriorates, the area near the peak appearing in the frequency spectrum of the beat signal expands accordingly. In other words, as the noise component included in the oscillation signal increases, the base of the peak expands in the frequency axis direction. Due to this effect, a number of small peaks due to noise appear near a peak having a large peak value, which may be erroneously recognized as a target peak.

Therefore, as in the case of the first embodiment, a level exceeding the maximum value of the level obtained by synthesizing the level of the foot portion near the peak due to the CZN characteristic and the random noise (the threshold value of the noise is a predetermined value) If the probability exceeds, set the value.

[0059] Fig. 11 shows a state of a frequency spectrum generated by combining a spectrum spread of a foot portion near a peak due to the CZN characteristic of the oscillator and a knock ground noise.

If a peak is detected based on only the value TH1 based on the background noise shown in FIG. 11, a plurality of peaks indicated by a plurality of circles included in a portion surrounded by a broken line A in FIG. It is erroneously detected as one get peak.

Thus, as shown in FIG. 12, the value TH3 is determined based on the peak value of the peak P and the CZN characteristic of the oscillator, and the value TH3 is determined based on the background noise. High! ヽ is used as the overall threshold. In this way, it is possible to prevent a peak caused by random noise from being superimposed on the foot near the peak due to the noise superimposed on the oscillation signal from being erroneously detected as the target peak.

[0062] As in the case of the first embodiment, such setting of the threshold value first detects a plurality of peaks exceeding a threshold value TH1 determined based on the background noise, and sets a plurality of peaks. The peak value of the peak and the power of the peak are sequentially increased, and the process may be repeated until there is no peak exceeding the threshold value.

Next, a radar according to a third embodiment will be described with reference to FIG.

The IF amplifier 7 shown in Fig. 1 is determined according to the distance of the beat signal to the target. It has the characteristic of changing the degree of amplification depending on the complete frequency. The IF amplifier circuit 7 increases the degree of amplification of the received signal as the signal is reflected from a distance, that is, as the frequency is higher. Therefore, the knock ground noise tends to increase as the frequency increases.

FIG. 13 shows an example of a frequency spectrum in that case. In this example, a peak P occurs in the range bin 31, and at the lower end, the base portion attenuates relatively sharply! /, And at the higher end, high-noise noise appears. Therefore, the threshold value determined based on the peak value of the peak P is set to be relatively abruptly attenuated on the low frequency side as shown by the threshold line T H2L in consideration of the characteristics of the distance attenuation correction described above. As shown by the value line TH2H, the attenuation is moderate on the high frequency side. However, in the example of Fig. 13, the threshold line TH2H on the high frequency side is approximately constant (slope 0).

As in the case of the first embodiment, a plurality of peaks exceeding a threshold value TH1 determined based on the background noise are detected, and the plurality of peaks are detected. The peak value of the peaks and the power of the peaks are sequentially increased, and the process may be repeated until there is no peak exceeding the threshold value.

Next, a radar according to the fourth embodiment will be described with reference to FIGS.

In FM-CW radar, signals from the switching power supply in the radar device or the clock oscillator of the signal processing circuit's beam scan mechanism may be mixed into the receiving stage. As a result, the beat signal contains FM or AM modulation components in addition to the original beat signal components of the reflected wave of the target and the transmission signal, and the side band components appear as spurious components in the beat signal. If the peak value of the original target peak is high !, the peak of the sideband will be a threshold even if the FM or AM modulation component is suppressed to a level that is smaller than the level of the original beat signal. And it may be erroneously detected as a target peak.

[0066] FIGS. 14A and 14B show the case where an AM modulation component is present in the beat signal, and FIG. 14A shows the case where the SZN ratio is relatively small and the peak value of the target peak is relatively small. , (B) is the case where the SZN ratio is relatively large and the peak value of the target peak is relatively large As described above, when the peak values of the target peaks PI and P2 are relatively low and the intensity of the background noise is high, the threshold value TH1 determined based on the background noise is high, and the sideband component is low. Since the peak is small, the peak due to the side band does not exceed the threshold TH1, but as shown in (B), the peak values of the target peaks PI and P2 are high, and the intensity of the knock ground noise is low. In the case of low ヽ, a peak due to a side band indicated by a circle exceeding the threshold TH1 is erroneously detected in a portion indicated by A near the target peak.

However, since the spectrum of a signal source generated as noise in the apparatus is substantially constant, the occurrence position and intensity of a side band generated near the target peak due to the signal source should be predicted. Can be. Therefore, as shown in FIG. 15, when the peaks of these side bands do not exceed the values, the values are set near the detected peaks.

In FIG. 15, (A) is an example in which the threshold value of the frequency range in which a side band near the peak occurs is increased. That is, an intensity higher than the predicted sideband intensity according to the peak value of the detected peak P1 is set as the threshold TH41. Similarly, an intensity that is higher than the predicted sideband intensity according to the peak value of the peak P2 is set as the threshold TH42.

FIG. 15B shows an example in which the threshold value is increased only at the predicted sideband noise position.

[0069] In the same manner as in the first embodiment, such a threshold value is set by first detecting a plurality of peaks exceeding a threshold value TH1 determined based on the background noise, and setting the plurality of peaks. The peak value of the peak and the power of the peak are sequentially increased, and the process may be repeated until there is no peak exceeding the threshold value.

In this way, it is possible to prevent a noise peak due to a side band generated by a modulation component from being erroneously detected as a target peak.

Claims

The scope of the claims

[1] means for transmitting a frequency-modulated transmission signal, generating a beat signal including a frequency component of a difference between the frequency of the reflection signal from the target of the transmission signal and the frequency of the transmission signal, Means for obtaining a frequency spectrum, and means for obtaining a peak frequency of a peak appearing in the frequency spectrum, wherein the radar is configured to detect a target based on the peak frequency.

Based on the intensity of the background noise or the reflected signal intensity of the target having a predetermined reflection cross section, the first threshold is determined, and the first threshold appearing in the frequency spectrum, the plurality of values exceeding the value are determined. Means for determining a second threshold value in a predetermined frequency region near each peak in accordance with the intensity of each peak, and extracting a peak exceeding the second threshold value. Characteristic radar.

[2] The second threshold value is characterized in that the value is increased in a predetermined band of a foot portion in accordance with the spread of the peak in the frequency axis direction caused by multiplying the beat signal by a window function. The radar according to claim 1.

[3] The second threshold value is characterized in that the value is increased in a predetermined band of a foot portion according to the spread of the peak in the frequency axis direction caused by the CZN characteristic of the oscillator generating the transmission signal. The radar according to claim 1 or 2, wherein:

[4] The second threshold value according to claim 1, wherein the second threshold value is determined so as to gradually decrease in a vertical direction with the frequency of the peak as a center according to the intensity of the peak. A radar as described in 2 or 3.

5. The second threshold value according to claim 14, wherein the second threshold value is set so as to exceed the intensity of a side band appearing together with the peak due to a modulation component superimposed on the beat signal. A radar according to any of the above.

[6] The means for extracting the peaks determines the second threshold value in the descending order of the peak value among the plurality of peaks exceeding the first threshold value, The radar according to any one of claims 15 to 15, which extracts peaks that exceed all of the thresholds.